B2B B2C Architecture for Smart Meters using IoT and Machine Learning: a Brazilian Case Study

In this paper, fault-tolerant architecture with a focus on usability integrates Internet of Things, Mobile, Cloud, and Machine Learning technologies to smart meter devices. Based on a survey with end-users, we developed a smart meter with an integrated display, mobile application, and a cloud-based data lake. After deployment in four households with different profiles in Brazil, energy state labeling with Hidden Markov Models and daily energy prediction with Machine Learning algorithms are implemented based on collected data. Prediction and information extraction from smart meter data can create value for both businesses and consumers.